An apparatus, method and computer program product are provided to train a feature detector to identify a respective feature from images captured by a camera. With respect to an apparatus, the apparatus causes at least one feature from one or more images that have been labelled to be projected onto a map. The apparatus is also caused to refine a representation of a path of a vehicle that carries a camera that captured the one or more images based upon registration of the at least one feature that has been projected with the map. Based upon the path of the vehicle following refinement, the apparatus projects one or more other features that have not been labelled from the map into the one or more images and then utilizes the images to train a feature detector.
Legal claims defining the scope of protection, as filed with the USPTO.
1. An apparatus comprising at least one processor and at least one non-transitory memory including computer program code instructions, the computer program code instructions configured to, when executed, cause the apparatus to: cause at least one feature from one or more images that have been labelled to be projected onto a map; refine a representation of a path of a vehicle that carries a camera that captured the one or more images based upon registration of the at least one feature that has been projected onto the map; based upon the representation of the path of the vehicle following refinement, project one or more other features that have not been labelled from the map into the one or more images to automatically generate a label of the one or more other features; and train a feature detector to identify at least one of the one or more other features based upon the label automatically generated by the projection of the one or more other features from the map into the one or more other images.
2. An apparatus according to claim 1 wherein the computer program code instructions configured to, when executed, cause the apparatus to cause the at least one feature to be projected onto the map comprise computer program code instructions configured to, when executed, cause the apparatus to cause the at least one feature to be projected onto the map based upon a location at which the one or more images were captured.
3. An apparatus according to claim 1 wherein the computer program code instructions configured to, when executed, cause the apparatus to refine the representation of the path of the vehicle comprise computer program code instructions configured to, when executed, cause the apparatus to refine the representation of the path of the vehicle based upon the registration of the at least one feature from the one or more images that have been labeled to a corresponding feature represented by the map.
4. An apparatus according to claim 3 wherein the computer program code instructions configured to, when executed, cause the apparatus to refine the representation of the path of the vehicle comprise computer program code instructions configured to, when executed, cause the apparatus to modify a determined location of the vehicle upon capture of the one or more images based upon a difference between a location of the at least one feature as defined by the one or more images and a location of the corresponding feature represented by the map.
5. An apparatus according to claim 1 wherein the computer program code instructions configured to, when executed, cause the apparatus to project one or more other features comprise computer program code instructions configured to, when executed, cause the apparatus to project the one or more other features from the map into the one or more images based upon a location at which the one or more images were captured as modified based upon refinement of the representation of the path of the vehicle.
6. An apparatus according to claim 1 wherein the computer program code instructions configured to, when executed, cause the apparatus to project one or more other features comprise computer program code instructions configured to, when executed, cause the apparatus to exclude projection of a feature from the map into an image in an instance in which the projection of the feature overlaps with a dynamic object within the image.
7. An apparatus according to claim 1 wherein the computer program code instructions configured to, when executed, cause the apparatus to train the feature detector comprise computer program code instructions configured to, when executed, cause the apparatus to train each of a plurality of feature detectors, each feature detector configured to identify a different respective feature.
8. An apparatus according to claim 1 wherein the computer program code instructions are further configured to, when executed, cause the apparatus to receive images captured by a camera carried by a vehicle while in transit.
9. An apparatus according to claim 1 wherein the map comprises a map configured to support autonomous driving.
10. An apparatus comprising at least one processor and at least one non-transitory memory including computer program code instructions, the computer program code instructions configured to, when executed, cause the apparatus to: train a hyperspectral feature detector to identify at least one of a plurality of features based upon the at least one feature in one or more hyperspectral images of a field of view, wherein the at least one feature has been labeled; automatically label, with the hyperspectral feature detector, the at least one feature in a red green blue (RGB) image of at least a portion of the field of view; and train an RGB feature detector to identify the at least one feature based upon the RGB image in which the at least one feature has been automatically labeled by the hyperspectral feature detector.
11. An apparatus according to claim 10 wherein the computer program code instructions are further configured to, when executed, cause the apparatus to receive a plurality of hyperspectral images and a plurality of RGB images captured by a hyperspectral camera and an RGB camera, respectively, wherein the hyperspectral camera and the RGB camera are each carried by a vehicle while in transit.
12. An apparatus according to claim 10 wherein the computer program code instructions configured to, when executed, cause the apparatus to train the hyperspectral feature detector comprise computer program code instructions configured to, when executed, cause the apparatus to train each of a plurality of hyperspectral feature detectors, each hyperspectral feature detector configured to identify a different respective feature.
13. An apparatus according to claim 12 wherein the computer program code instructions configured to, when executed, cause the apparatus to train the RGB feature detector comprise computer program code instructions configured to, when executed, cause the apparatus to train each of a plurality of RGB feature detectors, each RGB feature detector configured to identify a different respective feature based upon the RGB image in which the respective feature has been automatically labeled by different one of the plurality of hyperspectral feature detectors.
14. An apparatus according to claim 10 wherein the one or more hyperspectral images that are labeled are a subset of the hyperspectral images that are captured.
15. An apparatus comprising at least one processor and at least one non-transitory memory including computer program code instructions, the computer program code instructions configured to, when executed, cause the apparatus to: generate an image mask of at least one feature from one or more hyperspectral images of a field of view; project the image mask onto a corresponding red green blue (RGB) image of at least a portion of the field of view; and train an RGB feature detector to identify the at least one feature based upon the RGB image onto which the image mask of the at least one feature has been projected.
16. An apparatus according to claim 15 wherein the computer program code instructions configured to, when executed, cause the apparatus to generate the image mask comprise computer program code instructions configured to, when executed, cause the apparatus to generate the image mask based on one or more unique spectral characteristics of the at least one feature.
17. An apparatus according to claim 15 wherein the computer program code instructions configured to, when executed, cause the apparatus to project the image mask onto the corresponding RGB image comprise computer program code instructions configured to, when executed, cause the apparatus to project the image mask onto the corresponding RGB image captured at a common location as the hyperspectral image from which the image mask is generated.
18. An apparatus according to claim 15 wherein the computer program code instructions configured to, when executed, cause the apparatus to project the image mask onto the corresponding RGB image comprise computer program code instructions configured to, when executed, cause the apparatus to label the at least one feature in the corresponding RGB image based upon projection of the image mask onto the corresponding RGB image.
19. An apparatus according to claim 15 wherein the computer program code instructions configured to, when executed, cause the apparatus to train the RGB feature detector comprise computer program code instructions configured to, when executed, cause the apparatus to train each of a plurality of RGB feature detectors, wherein each RGB feature detector is configured to identify a different respective feature based upon projection of a different image mask onto a corresponding RGB image, and wherein each image mask is generated for a different feature of a plurality of features.
20. An apparatus according to claim 15 wherein the computer program code instructions are further configured to, when executed, cause the apparatus to receive a plurality of hyperspectral images and a plurality of RGB images captured by a hyperspectral camera and an RGB camera, respectively, wherein the hyperspectral camera and the RGB camera are each carried by a vehicle while in transit.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
December 21, 2018
February 16, 2021
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.